Plant Protein Localization Using Discriminative and Frequent Partition-Based Subsequences

The function of proteins in the living cells varies with respect to their localizations. Extracellular plant proteins are responsible for vital functions such as nutrition acquisition, protection from pathogens, communication with other soil organisms, etc. Hence, characterizing these proteins and distinguishing them from intracellular proteins is of high interest to biologists. Nonetheless, the small number of available extracellular proteins for training makes classifying them difficult and challenging. This work focuses on distinguishing extracellular proteins using partition-based subsequences, i.e., subsequences of amino acids in special partitions within the protein sequences. The use of an associative classifier in this work helps to acquire a set of accurate, small and interpretable localization rules that can be used for further biological analysis. The achievement of 98.83% F-Measure for identifying extracellular proteins shows the appropriateness of the selected features and the classification method.